Presented at the 19th International Congress of Radiation Research, Aug. 25-29, 2019, Manchester UK (https://www.radres.org/mpage/ICRR2019Home) and 12th International Symposium on Chromosome Aberrations, Aug 27, 2019; Manchester, UK. Background: Biological dosimetry of ionizing radiation with the dicentric chromosome (DC) assay (DCA) is accurate for assessment of whole (WB) and partial body (PB) exposures. However, labor intensive analysis has significantly limited its application for large cohorts or for timely monitoring of exposures in therapeutic settings. We evaluate fully automated interpretation of the DCA using data from biodosimetry laboratories participating in IAEA Cooperative Research Project E35010. Aims: We determined the feasibility of: a) automating DCA interpretation, b) transferring and processing metaphase image data remotely, c) generating timely and accurate results, d) addressing interlaboratory variability in calibration curves and radiation dose estimation of samples of unknown exposure. Methods: The Automated Dicentric Chromosome Identifier and Dose Estimator (ADCI) uses image processing segmentation to find chromosomes and machine learning-based algorithms to recognize centromeres and designate DCs. ADCI corrects false positive DCs at the chromosome-level by eliminating incorrect detections, and at the image-level by selecting optimal metaphase cells from each sample. Derivation of calibration curves and dose estimations is then performed by ADCI software. Metaphase images for dose calibration and WB and PB radiation test samples from HC, CNL, LT-RPC, VN-DNRI, PHE, and THA were transferred to network storage in Canada and analyzed with ADCI for curve and dose accuracy and speed. Results/Conclusions: Transfers of 8,212 to 264,409 images from different labs required 12-24 hr, processing to identify DCs completed in 1-11 hr, and automated image selection, calibration curves and test sample dose estimates were performed in 1-3 days. Dose estimates fell within or close to the IAEA triage threshold (±0.5Gy) for WB exposures. Automated optimization of image selection produced accurate results, but pre-set model selection performed better for some laboratories. Dose estimates for 2-6 Gy PB exposures were within 2.5Gy and estimated irradiated cell fractions fell within the 95% confidence interval. The potential for automated DCA to improve dose estimation has been clearly demonstrated.